Software Fault Forecasting through Puffer-fish Optimization Algorithm-based Feature Selection Technique

Software Fault Forecasting (SFF) identifies faults earlier in the software advancement procedure. However, it is vital to address the issue of high dimensionality to achieve an effective fault forecasting model. Feature selection (FS) is an effective technique to identify the best features from a dataset. This study proposes an innovative FS approach, referred to as POAFS rooted on the Puffer-fish Optimization Algorithm (POA) to extract the best traits from a software dataset. POAFS draws inspiration from the puffer-fish’s defense strategy from the predator’s attack. The proposed POAFS approach was examined on five public NASA datasets using a four-supervised learning classifier, namely Naive Bayes, Support Vector Machine, K-nearest neighbour, and Decision Tree. The investigatory findings revealed that the prospective POAFS proposal substantially enhanced the veracity of the fault forecasting model besides achieving the superior features.